[R-sig-ME] post hocs for lmer / lme

[Mike Lawrence]

One approach I've been playing with lately is using bootstrapping &
LME to generate distributions of predicted values for each cell in the
design (using the full model on each iteration of the loop). These
distributions (or, more properly, the distributions of difference
scores for focal comparisons) can then be used for post hocs. I
haven't thought about application to nested designs though, so you may
have to be careful with how you bootstrap there.

On Wed, May 12, 2010 at 6:14 AM, Kay Cecil Cichini
<Kay.Cichini at uibk.ac.at> wrote:
> hello,
>
> i searched for a reference how to do post hocs for lmms or glmms but without
> success!
>
> i have a model with 2 nominal explanatory variables (2 levels and 4 levels)
> and 2 nested random variables. from my final model i want to examine certain
> combinations of the fixed factors.
> how is this best achived in LMMs / GLMMs?
> is it valid to re-order the levels of the factors and assess the
> t-statistics of the paramteres?
>
> any advise or pointers on useful references would be greatly appreciated!
>
> best,
> kay
>
> _______________________________________________
> R-sig-mixed-models at r-project.org mailing list
> https://stat.ethz.ch/mailman/listinfo/r-sig-mixed-models
>



-- 
Mike Lawrence
Graduate Student
Department of Psychology
Dalhousie University

Looking to arrange a meeting? Check my public calendar:
http://tr.im/mikes_public_calendar

~ Certainty is folly... I think. ~